The present invention relates to optical communication networks, and more particularly to a method and apparatus for monitoring the quality of optical channels of an optical link having polarization-mode dispersion (PMD) to intelligently determine and distinguish PMD problems in the optical link.
In a typical optical communications system, an optical signal in the form of a series of light pulses is emitted from a modulated optical transmitter comprising a laser diode. Each light pulse is of extremely short duration, such as 40 ps, and is roughly Gaussian shaped as a function of time. In the frequency domain, this signal comprises numerous frequency components spaced very closely about the nominal center frequency of the optical carrier, such as 193,000 GHz. As this type of modulated optical signal passes through an optical fiber, different frequency components of the optical signal travel at slightly different speeds due to an effect known as chromatic dispersion. In the course of the optical signal traveling through a very long fiber, such as fibers that are 200 km in length, chromatic dispersion causes a single pulse of light to broaden in the time domain, and causes adjacent pulses to overlap one another, interfering with accurate reception. Fortunately, many techniques are known for compensating for chromatic dispersion.
Another form of dispersion is becoming a limiting factor in optical communications systems as progressively higher data rates are attempted. Polarization-mode dispersion (PMD) arises due to birefringence in the optical fiber. This means that for two orthogonal directions of polarization, a given fiber can exhibit differing propagation speeds. A light pulse traveling through a fiber will probably, unless some control means are employed, have its energy partitioned into polarization components that travel at different speeds. As with chromatic dispersion, this speed difference causes pulse broadening and restricts the usable bandwidth of each optical carrier.
Schemes to actively compensate for PMD generally involve detecting the presence of polarization-dependent timing differences and either: a) applying delay elements to one or the other polarization to realign the timing of pulses; or b) controlling the polarization state of the optical signal upon entry into the fiber, or at intermediate points along the fiber, such that birefringent effects are minimized or canceled out. Active compensation techniques are required because the PMD of a given fiber varies over time due to temperature and pressure changes along the fiber and due to aging. A fiber that is installed above ground can exhibit fairly rapid fluctuations in PMD due to temperature and mechanical forces. A fiber buried underground can be sensitive to loads such as street traffic or construction work.
A modulated optical signal arriving at an optical receiver must be of sufficient quality to allow the receiver to clearly distinguish the on-and-off pattern of light pulses sent by the transmitter. Conventionally, a properly designed optical link can maintain a bit-error-rate (BER) of 10xe2x88x9213 or better. Noise, attenuation, and dispersion are a few of the impairments that can render an optical signal marginal or unusable at the receiver. Generally, when an optical channel degrades to a bit-error-rate of 10xe2x88x928, a communications system will automatically switch to an alternate optical channel that has a better BER.
One common method of analyzing the quality of a modulated optical signal is a so-called xe2x80x9ceye diagramxe2x80x9d, shown in FIG. 1. The eye diagram consists of overlaying successive frames of time-domain traces of the signal, with each frame corresponding to one period of the nominal periodicity of the modulation. As portrayed, the vertical axis represents instantaneous intensity of the received signal, and the horizontal axis corresponds to time. Many successive traces of transmitted xe2x80x9conesxe2x80x9d and xe2x80x9czerosxe2x80x9d define a region or window within the middle of the display. In the time axis, the window is bound on either side by the transitional leading and trailing edges of the pulses. Using this technique, a large clear area or xe2x80x9cwindowxe2x80x9d in the center with no encroachment from any side represents a good signal in that the presence or absence of a pulse during each clock cycle is clearly distinguishable.
Noise added to a signal appears as xe2x80x9cfuzzinessxe2x80x9d of the lines defining the window. Sufficient noise can even obliterate the appearance of the window, representing a bad signal in that xe2x80x9conesxe2x80x9d and xe2x80x9czerosxe2x80x9d are no longer distinguishable. Impairments in the time axis, such as chromatic dispersion or polarization mode dispersion, cause the transitional areas of the display to close in upon the window from either side. Overlapping of pulses can require more stringent synchronization of the receiver""s decision point, or even render the signal unusable.
A given optical receiver will automatically adapt to receive a modulated optical signal. Automatic gain control (AGC), frequency control, and phase lock-in are typically applied in sequence so that a threshold decision circuit can best sample the signal and decode every pulse. Superimposed upon the eye-diagram, an optimal point of operation for a threshold decision circuit would intuitively be at the center of the window, as shown by the xe2x80x9c+xe2x80x9d in FIG. 1. This means that the intensity threshold is about halfway between the zero values and one values observed on average.
Timewise, the center of the window corresponds to sampling the pulses in the middle of their duration when they tend to be of maximum intensity also shown by the xe2x80x9c+xe2x80x9d in FIG. 1. Intuitively, one can see how this choice for an operating point would be the most robust against either noise or timing impairments, which cause the window to shrink.
A received optical signal can undergo some degree of change in, for example, pulse width without having an immediate impact on the BER, as observed by this optimally positioned main decision circuit. A particular type of receiver has been developed having at least two independent decision circuits of the type just described. Reference is made to an article entitled xe2x80x9cQ-factor Measurement for High Speed Optical Transmission Systemsxe2x80x9d, authored by A. J. Ramos which is from proceedings of the SubOptic ""97 conference, San Francisco Calif. 1997,891, as well as an article entitled xe2x80x9cMargin Measurement in Optical Amplifier Systemsxe2x80x9d authored by Bergano, Kerfort and Davidson, Photonics Technology Letters, 5(1993)304, the teachings of both are incorporated herein by reference. A main decision circuit is dedicated to actual communications reception and is maintained at the optimum point, once it is established, within the center of the window. But for analyzing signal quality to a finer degree and for measuring degradation before it impacts the BER of the actual communications, an auxiliary decision circuit is used to probe the extents of the operating window. Robustness to timewise disturbances is gauged by directing the auxiliary decision circuit to sample at various time offsets relative to the optimum point. Findings by the auxiliary circuit may even be used to fine-tune the optimum decision point settings of the main decision circuit.
The auxiliary decision circuit is set to a given timing offset and its output is monitored for BER, especially in comparison to the output of the main decision circuit. The BER measurement at each operating point can typically take several minutes. Gradually, BER data is accumulated for every offset value. As expected, a plot of this data resembles an inverted Gaussian curve with a minimum BER occurring some optimum offset, as shown in FIG. 2. A similar plot is derived by varying the amplitude threshold of the auxiliary decision circuit.
All of this BER data may be summarized into a xe2x80x9cQxe2x80x9d factor or quality metric for the received signal. In general terms, the broader the range of timings over which a low BER can be sustained, the greater the Q factor of the signal. A receiver with an auxiliary decision circuit can measure and output such a Q factor.
The Q measurement is particularly useful for assessing and fine-tuning an optical path that is already operating at a healthy low BER. The Q measurement estimates a BER without requiring any actual bit errors to occur. A Q measurement covering the BER range of 10xe2x88x9213 to 10xe2x88x9220 may be completed in a few minutes, whereas an actual errored bit might not be observed for hours, days, or months.
When an optical path degrades, some corrective action may be necessary either to improve the optical path or to divert the communications traffic along an alternate channel or path that will work better. Yet, it is equally important to the integrity of the traffic bearing signal to avoid taking unnecessary corrective actions. Each adjustment or switching operation can temporarily disrupt the traffic-bearing signal.
Where adaptive PMD controls are employed along an optical link, PMD phenomena detected by instruments, such as PMD compensators, must be interpreted as being either attributable to perturbation of fiber characteristics at large or by failure of PMD equipment along the link. Making this distinction avoids taking unnecessary control or maintenance actions upon PMD-related controllers and compensators.
Therefore, a method and system are required for monitoring the quality of optical channels in an optical link to determine and distinguish PMD problems in the optical link across all the optical channels from PMD equipment failures along the link to avoid taking unnecessary control or maintenance actions upon PMD controller and compensators.
Accordingly, the present invention provides a method and apparatus that achieves technical advantages by providing PMD fault isolation along an optical link by correlating PMD notifications across multiple optical channels to determine if detected shifts of the PMD are attributable to bulk fiber disturbances or per-channel PMD malfunctions. The present invention provides this distinction to temper corrective actions that would normally be taken to correct PMD problems. When it is determined that the PMD of all optical channels suddenly shift, it is determined that this PMD shift is attributed to bulk fiber disturbances. On the other hand, when it is determined that the PMD of only one optical channel has shifted, then it is determined that the PMD problems are related to the associated optical channel and, thus, actions can be taken to correct the PMD problem in this channel. Making these distinctions avoids taking unnecessary control or maintenance actions upon related PMD controllers and PMD compensators.
According to a preferred method of the present invention, the method comprises performing fault management in an optical communications link having a plurality of optical channels, whereby each optical channel has associated therewith at least one PMD detector, such as a PMD compensator, generating an output indicative of PMD of the respective optical channel at the PMD detector. The method comprises the steps of obtaining a first output from a first PMD detector associated with a first optical channel, obtaining a second output from a second PMD detector associated with a second optical channel, and correlating the first output and the second output to determine a fault of the optical communications link and generate a fault signal indicative of the fault. The PMD detector outputs are correlated across multiple optical channels to determine a fault of the optical communications link, such as to determine if a PMD problem is attributed to bulk fiber disturbances, or isolated PMD detector malfunctions.
The apparatus comprises an optical fiber, an optical transmitter coupled to the optical fiber for generating a first optical channel and a second optical channel through the optical fiber, a first PMD detector and a second PMD detector, such as a PMD compensator, coupled to the optical fiber for generating a first output and a second output, respectively, indicative of PMD of the first optical channel and the second optical channel, respectively, at a first location and a fault detector coupled to the first PMD detector and the second PMD detector for generating a fault output as a function of the first output and the second output indicative of a PMD problem in the optical fiber.
In summary, the present invention provides adaptive PMD controls, such as a PMDC, to be employed along an optical fiber link a fiber, whereby PMD phenomena detected by instruments, such as PMD compensators, can be interpreted as being either attributable to perturbation of fiber characteristics at large or by failure of PMD equipment along the optical fiber link or fiber. Making this distinction avoids taking unnecessary control or maintenance actions upon PMD related controllers and compensators.